Ethnic diversity in the UK: new opportunities and changing constraints

ABSTRACT The study of ethnicity and migration is a fast evolving field. Much remains to be understood about economic, social, demographic and health outcomes across ethnic groups and generations, and the extent of integration and exclusion of different immigrant-origin groups in countries of destination. While cross-national studies are increasingly shedding light on issues of differential migrant selection and institutional influences, detailed single-country studies, based on high quality nationally representative data have much to offer in enhancing our understanding of diversity within and between groups. The introduction to this special issue on ‘ethnic diversity in the UK: new opportunities and changing constraints’ provides a background to the contributions in the issue by outlining key research agendas relating to immigrant and ethnic minority groups in Western Europe. It then elaborates why the UK represents a good case for exploring ethnic diversity and change; and outlines the contributions made by the papers in the issue. We highlight the distinctive features of the data source used in the papers, Understanding Society, the UK Household Longitudinal Study. We also consider limitations in the data and in the applications in the issue. We conclude by briefly outlining the policy-relevant contributions of the papers in the issue.


Introduction
The aim of this special issue is to bring together papers from multiple disciplines that build on and extend existing theoretical and empirical developments, utilising nationally representative data and a range of quantitative analyses. Focusing on one country, and thereby holding the national context constant, this volume illuminates the contemporary experience of minorities across generations and across a series of distinct, but often interlinked domains: occupation and labour market, youth transitions, attitudes, identity, politics and health. The focus of this special issue is the UK as it is ideally suited to study different facets of the experience of migrants and ethnic minorities both due to the heterogeneity within its immigrant and ethnic minority population as well as the availability of data from a longitudinal household survey conducted on a nationally representative sample and ethnic minority and immigrant boost samples. All the papers utilise this single data source, Understanding Society: the UK Household Longitudinal Study (UKHLS).
The papers in this issue offer insights into contemporary economic position, the dynamics of ethnic minorities' experiences, and explore areas that are becoming a greater focus of interest in quantitative, multidisciplinary migration studies. They draw on theory, methods and approaches from sociology, economics, demography and political science. Together and separately, the papers have the potential to move forward the evidence base and, in some cases, the theoretical debate as well. They also offer some valuable evidence for policy-making.

Current directions in research on migration and ethnicity
The study of ethnicity and migration is a fast evolving field. Alongside current interest in grasping the transnational (Waldinger 2015) and cross-national features of migration (Crul and Schneider 2010;Diehl et al. 2016), there remains much to be learnt about those of immigrant origin within national settings. This includes diversity of experience within and across groups and over generations, and the implications for the future in countries of settlement.
Concerns with migrantand second generation -'integration' remain highly salient for both researchers and policy makers (Dustmann, Frattini and Lanzara 2012;Alba and Foner 2015;Diehl et al. 2016;Casey 2016). These concerns are enhanced by the perceived migration 'crisis' (Holmes and Castañeda 2016), high levels of anti-immigrant attitudes across a range of countries (Czaika and Di Lillo 2017), increasing support for far right parties and the rise of Islamophobia (Massoumi, Mills, and Miller 2017), and concerns about terrorism and cultural change (Card, Dustmann, and Preston 2012). Contemporary research on specific countries has the potential to inform these salient debates, and shed additional light on these phenomena and their drivers. The weight of research on longer term outcomes of the 'second generation' grows, as increasing numbers of children and even grandchildrenof immigrants reach adulthood (Portes and Rumbaut 2001;Heath, Rothon, and Kilpi 2008;Crul, Schneider, and Lelie 2012). Yet the outcomes and experiences of the descendants of immigrants are varied and complex; and our knowledge is subject to ongoing revision. As younger cohorts reach adulthood, as analysis more fully incorporates relevant background factors (Dustmann, Frattini and Lanzara 2012;Zuccotti 2015;Li and Heath 2016), and as the institutional and economic context itself changes (McGinnity et al. 2018), we develop fresh understandings of how minority groups are faring. This fresh understanding in turn indicates the extent to which 'integration' and anti-discrimination policies have been successful and the directions for future policy.
We have learnt more about how immigrants respond to shifts in migration rules and regimes, as well as to institutional characteristics of context and labour market conditions (Kogan 2006;Ersanilli and Koopmans 2010;De Haas 2011;. Such research has enabled explicit testing and reformulation of theories of assimilation (Alba and Nee 1997), segmented assimilation (Portes and Zhou 1993), and ethnic revival (Rumbaut 2008). At the same time differences in migrant selection and in context of reception in different environments (Portes and Rumbaut 2001;Dustmann and Weiss 2007;Crul and Schneider 2010;Haberfeld et al. 2011;Ichou 2014;Luthra, Soehl, and Waldinger 2017) may argue for the need to enhance understanding of the empirical realities and changing processes in a single context. Such a focus on a single country context can shed light on how outcomes across groups are (differentially) shaped by multiple factors and across multiple domainsfrom education (e.g. Kristen and Granato 2007;Heath and Brinbaum 2014;Strand 2014;Fernández-Reino 2016), to income and wealth (Fisher and Nandi 2015;Shapiro 2017), to health and mortality (Scott and Timaeus 2013;Wallace and Kulu 2015), to life satisfaction and belonging (e.g. Raijman and Geffen 2018). This is all the more the case given the speed and dynamics of change in migrant populations, and in their characteristics and outcomes. These changes imply further theoretical development embedded in careful contemporary analysis. Cross-sectional analyses are important in demonstrating contemporary experiences of minorities, comparing them with those of the majority, and relating them to individual and contextual factors. They are more limited in the extent to which they can disentangle period and cohort effects, as well as causal processes (Gayle and Lambert 2018). Dynamic analysis is therefore needed to shed light on how minorities' lives are changing. Yet there is still a relative paucity of data that facilitates such analysis for nationally representative minority group populations.
Extensive literature on 'ethnic penalties' (Heath and McMahon 1997;Heath and Cheung 2007) highlights the discrimination and disadvantage minorities face in the labour market, which is separately attested in a number of self-report studies (e.g. FRA 2017) and correspondence tests (Riach and Rich 2002;Booth, Leigh and Varganova 2012;McGinnity and Lunn 2011;Bursell 2014;Zschirnt and Ruedin 2016). But recent research also highlights the striking levels of educational success and upward social mobility that many minority groups achieve (Papademetriou, Sumption, and Somerville 2009;Strand 2014;Zuccotti 2015;Gracia, Vázquez-Quesada, and Van de Werfhorst 2016;Zuccotti and Platt 2017a). While elite and privileged migration has been studied as a domain in its own right (e.g. Beaverstock 2005;Favell 2011), the more complex interplay of privilege and disadvantage, selection, and achievement of immigrants and their children is ripe for further analysis.
Patterns of ethnic mobilisation are also changing. Traditionally supporters of left-wing parties, some minorities also demonstrate affiliation with more conservative political agendas (Dancygier 2013), potentially remaking political landscapes, especially as minorities become more socio-economically mobile. In a context of heightened sensitivity to migration, cultural cleavages and global movements and processes, issues of identity and group boundaries are contested and reclaimed (Reeskens and Wright 2013;Martinovic and Verkuyten 2014;Nandi and Platt 2015;Diehl et al. 2016). This has led to a re-envisioning of European migration and the integration models founded in analyses of former labour migrants (Favell 2008;Engbersen et al. 2013;Luthra, Platt, and Salamónska 2016).
The recognition of between-group and within-group diversity and change across generations that challenges linear accounts or generalisations (Portes and Rumbaut 2001), also benefits from the incorporation of group-level characteristics, namely socio-economic composition and geographical concentration (Borjas 1995;Urban 2009). Ethnic concentration has both potential positive and negative impacts on minorities' education, health and wellbeing, and labour market outcomes (Portes and Jensen 1989;van Kempen and Özüekren 1998;Gijsberts and Dagevos 2007;Xie and Gough 2011;Wessendorf 2017), but the mechanisms driving these associations for different groups merit further attention.
Attention to the relationship between 'home' and 'host' countries, whether in terms of cultural persistence or migrant selection (Fernandez and Fogli 2009;Kanas and Van Tubergen 2009;Haberfeld et al. 2011;Ichou 2014;Polavieja 2015;Finseraas and Kotsadam 2017) has enhanced our understanding of both group and individual-level experiences and outcomes. In addition, European minorities' experience cannot be divorced from wider national and economic trends such as the Great Recession and the 'migrant crisis' (Fisher and Nandi 2015;Holmes and Castañeda 2016). Qualitative literature has elaborated the complexities of ethnic minorities' and migrants' experiences of inclusion and exclusion (e.g. Baban, Ilcan, and Rygiel 2017), relationship with the state (e.g. Shutes and Walker 2018), transnational connections (e.g. Sperling 2013), multiplicity of motivations (e.g. Rutten and Verstappen 2014), and identities and belonging (Faas 2008;Wessendorf 2017), in ways that cannot be captured so readily in quantitative analyses. There remains, nevertheless, enormous scope for systematic quantitative analyses to complement such studies and to enhance our understanding of patterns of diversity and change across ethnic and immigrant-origin groups.

The UK case
The UK provides an interesting case for research on ethnicity and migration for many reasons. First, the experience of its main ethnic minority groups is marked by extensive diversity both across and within groups. Substantial advantage and extensive upward mobility co-exists with persistent disadvantage (Platt 2007;Longhi, Nicoletti, and Platt 2013;Strand 2014;Zuccotti 2015;Fernández-Reino 2016;Li and Heath 2016;Zuccotti and Platt 2017;Lessard-Phillips et al. 2018;Zwysen and Longhi 2018). Even among the stereotypical 'model minorities', such as British Chinese and Indians, research is now showing a more complex picture including substantial ethnic penalties among subgroups and at certain points in the economic distribution, linked in part to occupational concentration (Longhi, Nicoletti, and Platt 2013;Longhi and Brynin 2017;Mok and Platt 2018). Research on the UK's ethnic groups can thus shed light on factors that enable success alongside those that constrain opportunities.
Second, the UK has a history of differentiated migration flows. While in large part stimulated by strong labour demand, these have also been created by external circumstances such as the expulsion of East African Asians in the early 1970s and refugee flows stimulated by the Vietnam War and more recently the Syrian civil war. Following the expansion of the EU, migration from Eastern Europe (Burrell 2010) has changed both the scale and nature of immigration. These migration patterns along with changes in immigration law (McDowell 2009;Scott 2017), as well as global trends such as in the internationalisation of higher education (Findlay 2011; have shaped the contemporary composition of minorities. The UK therefore comprises an ethnically and socio-economically diverse landscape, which facilitates exploring similarities and differences in trajectories and outcomes. Third, this long history of migration has had two other consequences: multiple (migrant) generations and changes in the experiences of minorities over time as a consequence of the time spent in the UK and UK policies. For example, all minority ethnicities, including those who are seen as facing greater challenges in the education system (Burgess and Greaves 2013;Shaw et al. 2016), are now outstripping the white majority in their university attendance (Crawford and Greaves 2015). This raises questions about past assumptions relating to educational trajectories, as well as drawing attention to labour market exclusion of the highly qualified (Lessard-Phillips et al. 2018;Zwysen and Longhi 2018). The emergence of second and third generations, and multiple cohorts, now offers a strong basis for investigating second (and later) generation outcomes across multiple areas of life, and how they diverge from both the first generation and the majority population (Berrington, Roberts, and Tammes 2016;Luthra and Platt 2017).
Finally, the UK policy context is characterised by highly stringent immigration controls, alongside a dominant discourse of racial equality (Joppke 1999). Strong anti-immigrant attitudes and sentiments (Czaika and Di Lillo 2017) and increasingly restrictive migration policies (Scott 2017) are accompanied by historically relatively strong anti-discrimination legislation. At the same time, the relative flexibility of the labour market has been argued to be beneficial for migration (Kogan 2006). This provides a complex institutional context of both exclusion and inclusion.

Contribution of the special issue
The papers in this Special Issue demonstrate recent findings and between them, they illustrate the dynamics of experience across ethnic groups. A number of papers show engagement with new and emerging areas and approaches in migration research. We next treat these three features of the special issue, placing them in the context of existing research.

The relevance of current findings
Evidence from numerous countries illustrates how, as the second generation grows up and the first generation ages, alongside the arrival of new streams of migrants who differ from the main post-war labour migration flows, we may need to reconsider accepted wisdom on minority outcomes. The consequences of the changing demography of Europe's ethnic groups are reflected in increasing attention to the experience of aging among minorities (e.g. Moriarty and Butt 2004;Warnes et al. 2004). At the other end of the life cycle, life course patterns and events, including fertility as well as labour market transitions can now be researched among those who immigrated as children or the second generation (Kulu et al. 2017). Such research invites questions about expectations for family and the life course, and the extent to which these are informed by (competing) cultural expectations (Fernandez and Fogli 2009;Adsera and Ferrer 2014;Baykara-Krumme and Milewski 2017). Understanding differing expectations of family life across groups prior to the transition to adulthood can shed light on the subsequent interaction of labour market and partnership opportunities and constraints. In this issue, Berrington (2020) addresses this question of family formation expectations, and interrogates how far they are distinctive across those of different ethnicities raised in a given country and from similar age cohorts.
One of the key areas where we have seen change over time has been in the educational and occupational profiles of minority ethnic groups. Despite the emphasis in much of the European literature on understanding ethnic employment penalties and educational deficits (Heath and Cheung 2007;Heath and Brinbaum 2014), the picture is now changing rapidly, particularly in regard to education. The high educational aspirations of minorities are recognised across a range of contexts (Kao and Tienda 1998;Cebolla Boado 2011;Strand 2014;Alba and Foner 2015;Fernández-Reino 2016). While educational outcomes do not consistently map on to these aspirations, in some contexts educational performance is also high, with substantial variation across countries of destination among those from the same country of origin (Crul and Schneider 2010). The UK is one such context where, as noted, men and women both from traditionally high-performing groups and also those who had previously lagged behind are now outstripping the majority in university participation (Crawford and Greaves 2015;Boliver 2016). This raises rather particular issues for policy: instead of early 'compensation', it is in the transition from education to the labour market that the biggest challenges appear to arise. Moreover, the fact that social class background is less influential on educational outcomes for minorities, indicates potential policy learning in relation to social mobility.
Labour market performance, therefore, now needs to be understood in relation to higher as well as lower educational attainment, transforming the classic explanatory models applied to the (early) second generation across much of Europe. The labour market profile of ethnic minorities is itself showing some change as younger cohorts mature. While there is continuing differentiation in occupations across groups, particularly in relation to self-employment and 'niche' economies (Longhi and Brynin 2017;Mok and Platt 2018), there is also occupational upgrading and change (Zuccotti 2015;Li and Heath 2016;Zuccotti and Platt 2017). Both occupational patterns and access to the labour market in the UK continue to be partly shaped by the persistence of labour market discrimination (Booth, Leigh, and Varganova 2012;Bursell 2014;Zschirnt and Ruedin 2016). The complex patterning of labour market access and exclusion, which is also differentiated by gender (Zwysen and Longhi 2018), invites reconsideration of the role of occupational concentration as it plays out over time. This subject is explored by Longhi (2020) in this issue; while Li and Heath (2020) investigate the extent to which unemployment 'scarring' is ethnically differentiated. This can help to inform and update the theoretical arguments around the processes by which labour market disadvantage (and inclusion) arises; as well as emphasising the ongoing role for policy in tackling ethnic inequalities in the labour market.
In the UK, ethnic groups' composition is itself changing with changing immigration controls (Home Office 2014; Luthra and Platt 2017), changing migration trends following EU expansion (Luthra, Platt, and Salamónska 2016), and changing political context. Antiimmigration sentiment linked to the 2016 Brexit vote; and the far-reaching consequences of changes to migration policy since 2010, with an emphasis on deterrence, have created a 'hostile environment' even for long-standing minority populations (Parliamentary Statement, Hansard 2018), rendering boundaries 'brighter' (Alba 2005) and politicising identity claims. The ethnic majority is also changing, and experiences its own cleavages, with regional and country identities typically being felt more strongly than British or UKwide identities (Nandi and Platt 2015), and with ethnicity implicated in 'political' decisions (Henderson et al. 2017). Both proponents and critics of multiculturalism, have located such ethnic polarisation in the nature of immigration and integration policies and forms of reception (Koopmans 2013). In this issue, Nandi and Platt (2020) address such politics of identity across the majority and minority populations.

The dynamics of ethnicity and migration
Quantitative ethnicity and migration research is paying increasing attention to the dynamics of experience, both intergenerational transmission from parents to children (Papademetriou, Sumption, and Somerville 2009;Dustmann, Frattini and Lanzara 2012;Zuccotti 2015;Gracia, Vázquez-Quesada, and Van de Werfhorst 2016;Li and Heath 2016;Zuccotti and Platt 2017), and intra-individual dynamics, such as life course transitions and events, and their sensitivity to economic cycles (Fisher and Nandi 2015;Li and Heath 2016). Such longitudinal analysis both enables different questions to be answered than cross-sectional analyses (Bane and Ellwood 1986;Jenkins 2011) and makes a better case for causal relationships between variables. The use of longitudinal panel data enables greater attention to within-household relationships and intergenerational, parent-child transmission and change. While the amount of longitudinal analysis on ethnicity and migration across Europe is limited, its importance for addressing issues of migrant integration is increasingly recognised (Diehl et al. 2016).
Socialisation by parents and peers and intergenerational transmission of values and attitudes (De Hoon and Van Tubergen 2014) have been implicated in persistence of differences across groups over time. The role of 'culture' in helping to account for differences in women's fertility and labour market participation has received empirical attention in both economics (Fernandez and Fogli 2009) and sociology (Polavieja 2015). Such perspectives take us beyond standard assimilation accounts to consider how similar transmission processes can result in different patterns across groups. This can in turn prompt a reconsideration of the successes or failures of integration policy. In this issue, Arcarons (2020) addresses the issue of intergenerational transmission beyond the immediate family by considering the relationship between the participation of women's mothers-in-law and their own labour market attachment.
Household and intergenerational data can also shed light on socialisation and intergenerational transmission across other areas of life including specific attitudes, political affiliations and health behaviours. In their study of party affiliation among youth, Martin and Mellon (2020) take into account the potential influence of parents' political affiliations in shaping young people's political engagement. We know that health behaviours generally show strong intergenerational associations; and while Luthra, Nandi, and Benzeval (2020) do not address parent to child transmission directly, by focusing on the ethnic composition of networks and peer effects, their study relates to question of persistence of cultural transmission. Berrington (2020), meanwhile, compares attitude and expectations regarding family formation across cohorts within ethnic groups.

Emerging areas and unresolved questions
There are a number of areas in which multi-disciplinary quantitative research on ethnicity and migration using nationally representative data is evolving. Alongside a continued focus on educational and employment outcomes, more systematic attention has been paid within both sociology and economics to areas such as identity (e.g. Manning and Roy 2010;Fischer-Neumann 2014;Nandi and Platt 2015), attitudes (e.g. Khoudja and Fleischmann 2017), and life satisfaction (e.g. Bartram 2013) across ethnic groups. Research into the second generation also includes significant work on family and fertility, and political engagement (Kulu et al. 2017;Dancygier 2013;Heath et al. 2013). However, the substantial body of work on ethnic identity and intergroup relations has tended to neglect other aspects of identity that may inform minority and majority social interactions, and facilitate intergroup contact (Moody 2001). In this issue, Nandi and Platt (2020) evaluate the salience of, and factors shaping, political and ethnic identity for majority and minority groups. These identities may themselves have their roots in earlier direct political commitment and affiliations, the subject of Martin and Mellon's (2020) paper.
Despite a wealth of literature or ethnicity and health, the healthy migrant paradox (Abraido-Lanza et al. 1999), though a recognised phenomenon, is still contested in terms of its generalisability (Teruya and Bazargan-Hejazi 2013) and mechanisms. Recent attention has provided conflicting understandings of the evolution of health behaviours within and across minority groups. In this issue Luthra, Nandi, and Benzeval (2020) add to this literature by addressing the competing arguments and testing relevant mechanisms.
Studies of migration are increasingly sensitive to the potential and limits of existing data and methods for addressing outstanding questions and informing policy. For example, researchers now recognise the need to explore the distribution as well as 'average' experience (e.g. Longhi, Nicoletti, and Platt 2013), and to consider the relevant reference population for specific questions, which may be those in the origin country rather than the majority population at destination (e.g. Bartram 2013). A number of sophisticated analyses aim to take account empirically of the relevance of ethnic origins and processes of selection and cultural retention and transmission (Kanas and Van Tubergen 2009;Ichou 2014;Polavieja 2015;Finseraas and Kotsadam 2017). The papers in this special issue aim to draw on best methodological practice from across disciplines for the questions at hand, exploiting the potential of, but also within the constraints of, the data, discussed in the next section.

Understanding society: the UK household longitudinal study (UKHLS)
The UKHLS (University of Essex 2018) is uniquely suited for analysing and addressing research questions from multidisciplinary perspective for a number of reasons: its broad, nationally representative sample, its size and ethnic minority oversampling, longitudinal design; household level coverage; geographical spread; diverse content, and richness of measures of ethnicity, features that we now elaborate.
The UKHLS is a large-scale, nationally representative multi-topic household panel study with annual follow up (Knies 2017). As a longitudinal study, the original sample members are interviewed every year and asked a core set of questions, and some additional questions which are asked every few years. This makes it possible to analyse changes in people's lives, different events that occur and to follow up the consequences of earlier decisions, attitudes or behaviours. Because original sample members and their descendants are followed over time as long as they remain in the UK, the sample remains representative of the UK population over time.
The study started in 2009 with a general population sample (GPS) of around 26,000 and an ethnic minority boost sample (EMBS) of around 4,000 households. In 2015, an immigrant and ethnic minority boost sample (IEMBS) of 2,500 households was added (McFall, Nandi, and Platt 2017). The GPS was selected from the UK while the EMBS and IEMBS were drawn from Great Britain. These boost samples provide sufficient numbers from distinct ethnic categories to enable research about the UK's different ethnic groups and (migrant) generations. This feature is crucial to the potential for addressing the specific topics covered in this issue. Importantly, the UKHLS is the only longitudinal survey in the UK to enable analysis of changes in the lives of ethnic minorities. Articles in this issue exploit this longitudinal dimension analysing, for example, employment dynamics of adults (Li and Heath 2020;Longhi 2020) and party attachment acquisition among young people (Martin and Mellon 2020). The household focus of the study allows researchers to investigate the influence of family members on individuals' behaviour, attitudes and wellbeing. Almost all the articles in this special issue include household and family level characteristics such as household income or household size. Additionally, Martin and Mellon (2020) have included parents' self-reported party attachment to help understand the party attachment of youth; and Arcarons (2020) has exploited the fact that there is information from both partners/spouses in a couple, as well as retrospective data on parents, also used by Li and Heath (2020).
The wide geographical spread of the sample accompanied by the potential to link to external datasets at different geographical scales allows researchers to examine the role of area level characteristics on different outcomes. Such external data sets, include the decennial Censuses for the UK, area level employment and benefit data, the local area Indices of Multiple Deprivation (Noble et al. 2006), as well as more specialised data such as those on environmental air quality (Knight and Howley 2017) or housing markets (Bayrakdar and Coulter 2018). Articles in this issue (Luthra, Nandi, and Benzeval 2020;Nandi and Platt 2020) include area level characteristics, such as level of deprivation, and proportion of UKIP/BNP voters, in their analyses. The interaction between neighbourhood, often understood as community, inter-ethnic relations, and ethnic minority 'integration' has shown itself to be of enduring policy as well as academic interest (Finney and Simpson 2009;Casey 2016;HMG 2018).
Survey respondents are asked questions on a wide range of topicstheir family background, education, earnings and wealth, partnership and fertility behaviour, health, health behaviours and wellbeing, and attitudes and values. While most of this information is collected from 16+ year-olds, young people, 10 to 15 years old are asked to complete shorter questionnaires covering topics such as computer usage, bullying, and party attachment. As a result, the papers in this issue are able to analyse employment trajectories, health behaviour, political attachment, attitudes, preferences, and identities of adults and young people, giving a relatively full picture of experiences across ethnic groups. A sub-sample of respondents comprising mostly ethnic minorities, alongside a small comparison group of white UK adults, are asked additional questions relevant for ethnic minorities such as ethnic identity, harassment, discrimination, and migration history. These questions are analysed by some papers in the issue (e.g. Luthra, Nandi, and Benzeval 2020;Nandi and Platt 2020).
Survey respondents are also asked about their own ethnic group (census classification), own and (grand) parents' countries of birth and, for those not born in the UK, year of arrival in the UK, religious background and strength of identity. These measures, as well as collection of own ethnic group of different family members (spouses, children, parents), allow researchers to study the experiences of different ethnic (and ethno-religious) groups across generations and with sufficient flexibility to define the groups of interest for their research question. The role of ethnic categories and categorisation and the risks of essentialising the experience of minority groups has been much debated in the literature (see e.g. Brubaker and Cooper 2000;Jenkins 2014). While ethnicity is typically reported by respondents themselves, the constraints of the available categories can arguably send messages about key social divisions, racial thinking and hierarchies and may have limited connection with how people view their own self-identity (Ballard 1996;Kertzer and Arel 2000;Song 2004). It is therefore important that researchers have the opportunity not only to interrogate these categories; but also to formulate and define their own categories in relation to what boundaries (Alba 2005;Wimmer 2008) are meaningful for the question in hand, and for the points of reference in the literature, using the multiple measures available.
In this special issue, papers take different approaches to the construction of ethnic groups. For example, Li and Heath (2020), Longhi (2020), Nandi and Platt (2020), and Luthra, Nandi, and Benzeval (2020) use the 2011 census ethnic group categories to identify ethnic groups. They also split the groups into 1st generation if born outside the UK, 2nd and higher generation if born in the UK. Arcarons (2020) uses own, parents and grandparents' country of birth to construct ethnic / country of origin categories, and, since he is interested in the partnerships of his respondents, classifies those who arrived before age 14 with the second-generation, and those who arrived after that age as first generation. Berrington (2020) analyses 16-17 year olds who were born in the UK or who arrived in the UK prior to the age of 6 and categorises them into 2nd generation (where both parents were born outside UK) and the 2.5 & higher generation (where at least one parent was born in the UK). Martin and Mellon (2020), who also analyse a younger age group, 10-15 year olds, use ethnic group reported by these young people and their parents, where available, to measure their ethnic group. As their main focus is the period of socialisation of young people and these 10-15 year olds were socialised in the UK, they do not differentiate between generations. Rather than diversity across ethnic groups, Nandi and Platt's (2020) focus is on the minority-majority distinction itself and its implications for identity. Martin and Mellon (2020) also adopt this approach for the second part of their paper. Hence, while constructing groups on the basis of ethnic group (and country of birth), they focus on a binary distinction between those whose ethnicity is normalised as the majority and those who are 'minoritized' in ethno-national constructions (Faas 2008). The papers also reflect research domain and disciplinary boundaries in their terminology. The derivation of these categories is made explicit in the papers themselves.

Limitations
Before turning to the policy implications of the papers in this issue, we consider some limitations and areas for further development. These are both substantive and analytical.
Attention to heterogeneity within, as well as between groups is increasingly a feature of contemporary research and is an important consideration in relation to thinking about the relevance of categories as well as the appropriateness or not of policy responses based on 'average characteristics'. While the papers include some acknowledgement of heterogeneity, greater attention could be paid to dispersion or variation around the group average and ascertaining factors linked to such variation. In addition, while recognising that comparisons of minorities with the majority tends to normalise majority experience in ways that can limit analytical purchase as well as perpetuating the 'othering' of minorities, this continues to be the dominant approach for the majority of quantitative analyses of ethnicity, and this issue is no exception.
While the papers demonstrate some use of the dynamic and longitudinal aspects of the data, some are fundamentally cross-sectional analyses, even if bringing in retrospective data or information from different sweeps. This is partly an issue of the research questions selected, but also in part reflects the relatively short run of data waves available at the time of writing. Since not all questions are asked at every wave, the opportunity for even shortrun repeated measures analysis is limitedfor example, by wave 6 a number of modules had only been fielded twice and a few only once. Moreover, the benefit of longitudinal data that it follows the same people over timeis also a limitation when it comes to attrition. This is particularly the case for analysis of minority ethnic groups, since they both more likely to be lost to follow up (for a whole range of reasons), and are smaller in numbers, even in a boosted sample such as this, to start with. A related issue is that not all minority groups are covered in sufficient numbers in this survey to allow meaningful analysis. Examples might be some Eastern European migrant flows, or respondents from the 'third generation', which are of increasing research interest.

Policy implications of the contributions
The papers in this issue offer some valuable evidence for policy making. Li and Heath (2020) find that ethnic minorities in UK experience ethnic penalties in earnings and employment. They also experience worse scarring effects from past unemployment. Longhi (2020) finds that it is the duration of past unemployment and not the type of job they do that determines ethnic minorities' higher risk of losing a job. These findings imply that the point of intervention is when people are unemployed rather than changing the sorts of jobs that minorities go into; and point towards a focus on employers rather than the choices, preferences and behaviours of (potential) employees. Arcarons (2020) focuses on ethnic minority women's labour market participation, which has increasingly been the subject of debates, discussion and policy papers. His finding that the labour market behaviour of one generation of women impacts on the behaviour of the next generation of women, particularly for Pakistani and Bangladeshi women, highlights the importance and long run impact of policies that support women's labour market participation. Berrington (2020) finds similar family formation and partnership expectations across generations although second generation Indians report higher cohabitation expectations. This sheds some light on contemporary and future partnership and family patterns.
Decreasing party attachment among UK's youth has been a topic of recent policy discussions. But Martin and Mellon (2020) find that this is not true for ethnic minority youth. The reason they find this is that parents' party attachment is transmitted to their children and ethnic minority parents have stronger party attachment. Thus the key question for policy makers and political parties is not why young people are politically disengaged, but why ethnic minorities are more politically engaged. Nandi and Platt (2020) tackle a different question about UK politics: the extent to which 'politics of identity' is prevalent. They find that, even after controlling for individual factors, the broader contextual drivers of political and ethnic identities are positively correlated. They find this correlation to be stronger for the white majority, particularly those who express conservative or regional party affiliations. This sheds light on the way ethnonational identities are being constructed, with implications for electoral engagement.
Like Berrington and Martin and Mellon, Luthra, Nandi, and Benzeval (2020) examine cultural transmission of behaviour and find that minorities with greater ethnic attachment engage in healthier smoking and drinking behaviour. In other words, cultural factors provide health protection for ethnic minorities, an important corrective to policies focused on assimilative strategies.
Overall, and notwithstanding the limitations previously elaborated, the papers in the special issue provide an in-depth look into the lives and experiences of UK's ethnic minorities, across genders, generations and over time.